DeepL is taking a bigger swing at voice translation by buying Mixhalo, a San Francisco startup whose software delivers real-time audio to audiences at concerts, sports arenas and conferences. The acquisition gives the German translation company a way to move beyond documents and meetings and into one of the hardest environments for language technology: live, noisy, high-stakes events where attendees need instant access to what a speaker is saying.
The deal brings together two companies that had already been circling the same problem from different angles. DeepL has spent years building a reputation as one of the strongest machine translation providers for text, and more recently has been expanding into speech. Mixhalo, meanwhile, built a platform for routing live audio directly to attendees’ phones, first for music events and later for sports and enterprise use cases. Together, the companies are betting that event translation can be more seamless, more accurate and easier to deploy if audio delivery and language conversion sit under one roof.
For conferences, keynote talks and panel discussions, the need is obvious. Attendees who do not understand the speaker’s language often end up holding a phone close to a microphone, relying on a generic translation app and hoping the signal is clear enough to work. That approach is clumsy and often unreliable. DeepL and Mixhalo say the acquisition is meant to replace that patchwork experience with a purpose-built system that can stream translated audio in real time.
What DeepL is buying
Mixhalo was founded in 2016 by guitarist and songwriter Mike Einziger, violinist Ann Marie Simpson-Einziger and entrepreneur Vik Singh, who now runs the company as CEO. The startup initially focused on improving live music listening experiences, then broadened into sports and other live events, where it offered a way for audiences to hear a cleaner, more direct audio feed on their own devices.
Over time, Mixhalo built a business around low-latency audio delivery for large crowds. That technology is especially useful when event organizers want to provide multiple language tracks, alternative commentary feeds or accessibility options without changing the core venue setup.
The startup raised more than $39 million from investors including Fortress Investment, Founders Fund, Defy Partners and Cowboy Ventures. Its acquisition by DeepL marks a notable exit in the crowded event-tech and audio-streaming market, and it also highlights how translation companies are moving closer to the user interface rather than remaining behind the scenes.
Why live event translation is so difficult
Live translation is not the same as translating text after the fact. In a conference hall, a stadium or a concert venue, the system has to cope with background noise, interruptions, speaker accents, changing microphones and network variability. Any delay can make the translated output feel disconnected from the original speech. That is why the category has long been difficult for AI companies to master.
There is also a user-experience problem. Traditional translation tools were built for one person interacting with one device. Live events require one-to-many distribution, often to thousands of attendees at once. Organizers need tools that can be deployed quickly, work across devices and remain stable under heavy load.
Mixhalo’s technology addresses the distribution side of that equation by routing audio directly to attendees. DeepL adds the language layer. The combined proposition is that a user could receive translated speech in a way that feels closer to a broadcast feed than a mobile-app workaround.
DeepL’s move beyond text
DeepL built its brand on high-quality text translation and became a familiar name among businesses and knowledge workers that need reliable multilingual communication. But the company has been broadening its ambitions for several years.
In 2024, it introduced voice-to-text translation in more than 33 languages. In April 2026, it expanded again with voice-to-voice translation designed for use cases such as multilingual meetings. The Mixhalo acquisition extends that strategy into live public settings, giving DeepL a real-world showcase for its speech products.
DeepL CEO Jarek Kutylowski described Mixhalo both as a product and as a proof point. In his view, the platform can demonstrate DeepL’s real-time capabilities in a setting where the audience is physically present and the need for instant translation is easy to understand.
DeepL sees Mixhalo as more than a standalone business line. According to Kutylowski, it will also serve as a visible example of how the company’s real-time translation tech performs in conference halls and other in-person environments.
The acquisition also signals that DeepL wants to own more of the user journey. Instead of simply translating content, the company is trying to control how that content reaches people, especially in environments where timing and delivery matter just as much as accuracy.
How the deal came together
According to Singh, the discussions were not the result of a formal hunt for a buyer but grew out of a long-standing customer relationship. Mixhalo had already been using DeepL as its main translation provider, creating a natural overlap between the companies’ products and teams.
Singh said the acquisition talks developed organically. He explained that a customer dinner led to a conversation with DeepL’s CTO, Sebastian, and the more they spoke, the clearer the overlap became across event use cases, APIs and the broader application layer.
That overlap matters. DeepL brings language models, translation infrastructure and brand recognition in the translation space. Mixhalo brings a distribution layer, event expertise and a customer base already accustomed to real-time audio experiences. In a market where many AI vendors are racing to add voice capabilities, having both technology and a proven use case may give the combined company a stronger position.
Why the voice AI boom matters to Mixhalo
Singh said the explosion of voice models has been useful for Mixhalo because it gives the company flexibility. Different models can be plugged in and compared, allowing the startup to evaluate performance and choose the best option for a given scenario.
At the same time, he argued that the rise of larger model companies creates competitive pressure. As major AI players extend into adjacent products, smaller specialists can find themselves squeezed on price and distribution. In Singh’s view, that makes it harder to compete solely on cost.
The logic is familiar across the AI sector. When infrastructure providers expand their product scope, independent application companies can either differentiate strongly or risk being absorbed into a bigger ecosystem. In Mixhalo’s case, joining DeepL appears to provide both protection and a path to broader commercialization.
DeepL’s U.S. expansion plan
The acquisition is not only about product strategy. DeepL said the deal will also help it expand in the United States. Because Mixhalo is based in San Francisco, DeepL plans to open a Bay Area office as part of the acquisition.
That move gives the company a stronger foothold in one of the most important markets for enterprise AI, conferencing, events and multilingual collaboration tools. It also puts DeepL physically closer to many of the buyers, developers and event organizers it wants to serve.
For a European AI company trying to grow in the U.S., a local office is more than a symbol. It can help with sales, partnerships, recruiting and customer support, while also positioning the company within the broader Bay Area ecosystem of enterprise software and AI startups.
Where Mixhalo fits in the competition
Mixhalo is not the only company trying to solve live translation and event audio problems. DeepL identified competitors including Wordly AI and Palabra, the latter backed by Seven Seven Six. The category has attracted attention because multilingual events are common, but current solutions are still fragmented and often require specialized setup.
Competitors in this space generally fall into a few buckets:
- AI translation tools focused on meetings and webinars
- Event platforms that provide audio distribution but not language conversion
- Broadcast or captioning services adapted for live events
- General-purpose translation apps used as a stopgap by attendees
DeepL’s advantage, if the acquisition works as intended, is the combination of high-performing translation and a purpose-built live delivery system. That could make the company more competitive in environments where organizers want a polished attendee experience instead of a DIY workaround.
Why conferences are a strategic beachhead
Conferences may look like a niche, but they are a useful proving ground for live translation technology. They are multilingual, they are time-sensitive and they often involve audiences willing to try new tools if those tools improve comprehension.
They also create a visible demo environment. When technology works in a keynote hall, it is easier to sell that technology to organizers of corporate meetings, training events, trade shows and eventually larger public venues.
That is why the acquisition could matter beyond the immediate event market. If DeepL can show that translated audio works smoothly in conferences, it may be able to extend into adjacent categories such as:
- Corporate all-hands meetings
- Training sessions and workshops
- Live product launches
- Sports broadcasts and arena commentary
- Accessible audio feeds for public venues
Comparison of the two companies
| Company | Founded | Core focus before the deal | How it fits in the combined strategy |
|---|---|---|---|
| DeepL | 2017 | Text translation; recently voice translation | Provides language models and translation quality |
| Mixhalo | 2016 | Real-time audio delivery for concerts, sports and live events | Provides live audience distribution and event use cases |
| Combined company | — | Live multilingual communication | Aims to deliver translated speech in conferences and other in-person settings |
The business case behind the acquisition
Acquisitions in AI are increasingly about distribution as much as capability. Many companies can access similar models or build comparable features, but fewer have direct access to end users in specific settings. DeepL’s purchase of Mixhalo appears designed to solve that problem by embedding the translation engine into a real application layer.
The business case works on several levels. First, it gives DeepL a compelling demonstration environment for its voice products. Second, it adds a new market segment with clear enterprise demand. Third, it brings in a company that already understands event operations, audio delivery and audience-facing product design.
For Mixhalo, joining a larger translation company may provide the scale needed to survive in a market where giant AI players are rapidly expanding their offerings. For DeepL, it offers a shortcut into a category that would otherwise require years of product development, partnerships and event-industry relationships.
What the acquisition says about the broader AI market
The deal reflects a wider shift in the AI industry. In the early phase of generative AI, many companies focused on building models and attracting users through standalone demos. Now, the market is moving toward practical deployment in specific industries and environments.
Translation is a strong example of that transition. The technology is no longer just about converting text on a screen. It is increasingly about integrating into meetings, customer support, public events and cross-border communications, where reliability and speed are essential.
That means companies that control both the AI layer and the workflow layer may have an advantage. Rather than competing only on accuracy metrics, they can compete on the entire experience: setup, latency, device compatibility, speaker handling and audience satisfaction.
DeepL’s purchase of Mixhalo places it squarely in that race.
What happens next
DeepL has not disclosed financial terms of the acquisition. What is clear is the company plans to use Mixhalo as both a product line and a customer-facing showcase. The Bay Area office suggests DeepL is also serious about building a larger U.S. presence around the deal.
The immediate challenge will be integration. DeepL will need to combine Mixhalo’s live-audio infrastructure with its own translation systems while preserving the low-latency experience that event organizers expect. It will also need to make the product easy enough for conferences and venues to adopt without a major technical overhaul.
If the integration succeeds, the combined offering could become one of the more interesting examples of AI moving from a general-purpose utility to a specialized operational tool. Instead of asking attendees to fend for themselves with a translation app, event organizers could offer a native multilingual experience from the start.
For a company known primarily for text translation, that would be a meaningful shift. DeepL would not just be translating language anymore. It would be shaping how live audiences hear it.
Key facts at a glance
| Item | Details |
|---|---|
| Acquirer | DeepL |
| Target | Mixhalo |
| Founded | Mixhalo founded in 2016 |
| Founders | Mike Einziger, Ann Marie Simpson-Einziger, Vik Singh |
| Known funding | More than $39 million |
| DeepL voice milestone | Voice-to-text in 33+ languages in 2024; voice-to-voice launched in April 2026 |
| Strategic outcome | Expansion into live events, conferences and U.S. operations |
Why it matters
At first glance, the acquisition may look like a narrow product move. In reality, it illustrates how AI companies are trying to own entire use cases instead of isolated features. DeepL is using Mixhalo to move into a real-world setting where translation is both visible and valuable. Mixhalo, in turn, gets access to a larger distribution channel and a broader AI platform.
That makes the deal important not only for the two companies involved, but for the wider translation and voice-AI market. The next phase of competition is likely to be won by companies that can turn model performance into everyday utility. In live events, where the payoff is immediate and the limitations are obvious, that test may be especially revealing.
For now, DeepL has made its bet clear: the future of translation is not just on the page or in the meeting room. It is also in the crowd, in the venue and in the ear of every attendee who wants to understand what is being said in real time.









